[Freesurfer] subcortical masks
Hi, I'd like to create binary masks in the nifti-format from each of the subcortical segmentations in freesurfer, separately (one for hippocampus, one for thalamus, and so forth..). Is there any way to do this automatically from the command line? Thanks, Martin Ystad Ph.D. candidate, University of Bergen ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Version re-run
Hi all, I have 100-something scans already run on a FreeSurfer version from 2005 that I'd like to re-run with the newest version of FreeSurfer. The subjects contain manual edits to the brainmask.mgz and wm.mgz, as well as control points. As far as I have understood, when running autorecon2 my manual edits will be saved by default. However, is there any need to re-run the autorecon1-steps for my subjects? Also: when do you plan to release the next version of FreeSurfer? Many thanks, Martin Ystad ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] create volumes from surfaces
Hi, I'm working on a fractal dimension analysis of the pial- and white-matter surface. The matlab script that calculates the fractal dimension needs binary nifti-files as input. Since FreeSurfer creates the surfaces I need, here's what I want to do: I'd like to convert the pial and white-matter surfaces (separately) to .nii-volumes (e.g. 256x256) where the voxels intersected by the respective surfaces are white, and everything else is black. Furthermore, I'd like to do this for individual subregions of each surface; for example only the precentral gyrus or the whole frontal lobe, so that only the voxels intersected by the surface belonging to the specific label I choose appear white, and everything else is black. How do I get this output from FreeSurfer? Thanks, Martin Ystad PhD-candidate, University of Bergen, Norway ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Re: [Freesurfer] create volumes from surfaces
I tried this, but something is going wrong. Here's what I did: 1. I ran $ mri_binarize --i subjid/mri/aparc+asec.mgz --match xx --match yy --o lobe.mgz - this produces a nice binary mask of, say, the left frontal lobe. 2. Then I ran $ mri_surf2vol --mkmask --surf pial --hemi lh --template lobe.mgz --identity subjid --o lobe_surface.mgz - this produces a nice binary image of the pial surface, but it contains the whole surface, not only the intersection with lobe.mgz. What am I doing wrong? Thanks, Martin Douglas N Greve wrote: > Use mri_surf2vol to create volumes. You can binarize aparc+aseg.mgz to > give a mask of an area (use --match), then take the intersection of the > mask with the surface-in-volume. > > doug > > Martin Ystad wrote: > >> Hi, >> I'm working on a fractal dimension analysis of the pial- and >> white-matter surface. The matlab script that calculates the fractal >> dimension needs binary nifti-files as input. Since FreeSurfer creates >> the surfaces I need, here's what I want to do: >> I'd like to convert the pial and white-matter surfaces (separately) to >> .nii-volumes (e.g. 256x256) where the voxels intersected by the >> respective surfaces are white, and everything else is black. >> Furthermore, I'd like to do this for individual subregions of each >> surface; for example only the precentral gyrus or the whole frontal >> lobe, so that only the voxels intersected by the surface belonging to >> the specific label I choose appear white, and everything else is black. >> How do I get this output from FreeSurfer? >> >> Thanks, >> Martin Ystad >> PhD-candidate, >> University of Bergen, Norway >> ___ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu >> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >> >> >> >> > > ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Re: [Freesurfer] create volumes from surfaces
Hi, thanks for the answer. Sorry for my delayed response. I tried the mris_calc command, but it didn't work. Here's the error message I got: "mris_calc: curvature file 'lobe.mgz' has wrong magic number." The multiplication was easily done i Matlab, but is there any way to do this with FreeSurfer? Thanks, Martin Douglas N Greve wrote: > The 2nd step only creates a mask of the entire surface. You need a > step 3 in which you multiply them together: > > > 3. mris_calc -o product.mgz lobe.mgz mul lobe_surface.mgz > > Martin Ystad wrote: >> I tried this, but something is going wrong. Here's what I did: >> 1. I ran $ mri_binarize --i subjid/mri/aparc+asec.mgz --match xx >> --match yy --o lobe.mgz >> - this produces a nice binary mask of, say, the left frontal lobe. >> 2. Then I ran $ mri_surf2vol --mkmask --surf pial --hemi lh >> --template lobe.mgz --identity subjid --o lobe_surface.mgz >> - this produces a nice binary image of the pial surface, but it >> contains the whole surface, not only the intersection with lobe.mgz. >> What am I doing wrong? >> >> Thanks, >> Martin >> >> Douglas N Greve wrote: >>> Use mri_surf2vol to create volumes. You can binarize aparc+aseg.mgz >>> to give a mask of an area (use --match), then take the intersection >>> of the mask with the surface-in-volume. >>> >>> doug >>> >>> Martin Ystad wrote: >>> >>>> Hi, >>>> I'm working on a fractal dimension analysis of the pial- and >>>> white-matter surface. The matlab script that calculates the fractal >>>> dimension needs binary nifti-files as input. Since FreeSurfer >>>> creates the surfaces I need, here's what I want to do: >>>> I'd like to convert the pial and white-matter surfaces (separately) >>>> to .nii-volumes (e.g. 256x256) where the voxels intersected by the >>>> respective surfaces are white, and everything else is black. >>>> Furthermore, I'd like to do this for individual subregions of each >>>> surface; for example only the precentral gyrus or the whole frontal >>>> lobe, so that only the voxels intersected by the surface belonging >>>> to the specific label I choose appear white, and everything else is >>>> black. >>>> How do I get this output from FreeSurfer? >>>> >>>> Thanks, >>>> Martin Ystad >>>> PhD-candidate, >>>> University of Bergen, Norway >>>> ___ >>>> Freesurfer mailing list >>>> Freesurfer@nmr.mgh.harvard.edu >>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>> >>>> >>>> >>> >>> >> >> > ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Problems with FSGD file
Hi, I'd like to try out some simple regression analysis on some of my data. I've completed the tutorial until the mri_glmfit step where I get the error message: ERROR: gdfRead: gruppefil.txt is not formated properly I really can't see what's wrong with my file, I've even copy-pasted one of the example files from the tutorial just to check if anything is working at all, but it didn't work either. GroupDescriptorFile 1 Title KognitivAldringTEST21 Class Class1 CLASS Class2 Variables Age Input fs568 Class1 54 Input fs569 Class2 59 Input fs572 Class2 67 Input fs573 Class1 46 Input fs574 Class1 52 Input fs577 Class1 46 Input fs578 Class2 68 Input fs579 Class1 47 Input fs580 Class2 74 Input fs583 Class2 61 Input fs585 Class2 62 Input fs587 Class2 56 Input fs588 Class2 59 Input fs592 Class1 49 Input fs593 Class2 56 Input fs594 Class2 58 Input fs597 Class2 61 Input fs599 Class1 46 Input fs601 Class2 65 Input fs603 Class2 68 Input fs608 Class2 70 DefaultVariable Age I've tried with classes, without classes, with more variables and so on, but nothing seems to work. Any ideas on why I get this error message? I used emacs, vim and open-office to create my FSGD file, gruppefil.txt. It's probably just a trivial mistake somewhere, but I just can't seem to find it... Thanks, MartinYstad University of Bergen, Norway ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Resuming recon-all
Hi, some of my subjects crash due to memory-failure way into the recon-all stream. Is there any way I can resume the processing stream from where it crashed without having to write all the individual commands for the rest of the stream? For example, if my subject crashed during autorecon2 -calabel, would there be a command that re-runs -calabel and the rest of the autorecon2 stream? Thanks, Martin Ystad University of Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Version problems
Hi, I have currently processed some 100 brains in freesurfer on different machines. Unfortunately, one machine was installed with version 3.0.3 (64 bit version), while the other machines were using version 3.0.1 (32 bit). It now looks like there is a systematic difference in the results from the various versions. I tested my suspicion by processing several identical volumes in both versions. The 3.0.3 version under-estimated the Intra Cranial Volume compared to the 3.0.1 version (around 75% of the 3.0.1-volume). Most other measures were slightly smaller, but no more than between 95-99% of the 3.0.1-volumes. What is the reason for this? How much of the processing stream do I have to re-run? Thanks, Martin Ystad Medical Student University of Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Subcortical segmentation
Hi, I noticed that routinely checking the subcortical segmentation (aseg.mgz) is not part of the Recommended Reconstruction Work Flow. Is this because the subcortical segmentation is highly accurate, or because manual intervention in the aseg is not recommended anyway? Thanks, Martin Ystad PhD-student University of Bergen, Norway ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Re: [Freesurfer] Subcortical segmentation
Maybe I've misunderstood. In http://surfer.nmr.mgh.harvard.edu/fswiki/RecommendedReconstruction , after autorecon2 it says to check white and pial surfaces before running autorecon3. The linked tutorials on how to do WM, Pial and Final surfaces edits, does not mention checking the aseg as far as I can see. Also, to express myself more clearly: I'm thinking of the labeling of the subcortical structures (thalamus, putamen, etc.). Is there something I'm missing here? - Martin Allison Stevens wrote: > It should be part of the workflow. Can you show me where it was not listed? > > On May 13, 2010, at 7:58 AM, Martin Ystad wrote: > > >> Hi, I noticed that routinely checking the subcortical segmentation >> (aseg.mgz) is not part of the Recommended Reconstruction Work Flow. Is >> this because the subcortical segmentation is highly accurate, or because >> manual intervention in the aseg is not recommended anyway? >> >> Thanks, >> Martin Ystad >> PhD-student >> University of Bergen, Norway >> ___ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu >> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >> >> >> > > > The information in this e-mail is intended only for the person to whom it is > addressed. If you believe this e-mail was sent to you in error and the e-mail > contains patient information, please contact the Partners Compliance HelpLine > at > http://www.partners.org/complianceline . If the e-mail was sent to you in > error > but does not contain patient information, please contact the sender and > properly > dispose of the e-mail. > > ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Intensity problems
Hi, I'm currently running the latest developmental release for RedHat 9, but I seem to have some problems concerning the intesity-normalization. Unfortunately, my images have some intensity inhomogenities, but the first normalization-step seems to take care of most of these. However, some regions inside the white matter has a somewhat high value, for example 130 -140. These regions are not included in the white matter when I run the recon-all -segmentation program. Adding control-points to these areas does not work, obviously, since only values lower than 110 are considered when the normalization-step is re-runned. (right?) Is there any way to include these areas in my final white matter whithout manually drawing them onto the wm.mgz -volume? For instance raising the threshold for wm-segmentation? If so, would this produce any kind new artifacts which I need to be aware of? Thanks, Martin Ystad Medical Student University of Bergen Institute of Biomedicine Jonas Lies vei 91, 5009 Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Re: [Freesurfer] Intensity problems
Jenni Pacheco wrote: Hi Martin, Control points should fix the problem. Try adding some to this brighter region, as well as a few to regions currently labeled to 110. Sometimes this helps. After you've saved your control points (be sure to save them) then you should run: recon-all -normalization -usecontrolpoints -subjid to make sure that it uses your control points. If you look at your brain T1 volume now, those bright spots should be taken care of. If so you'll need to re-run these steps: recon-all -skullstrip -subcortseg -normalization2 -segmentation -subjid If it hasn't worked, let me know, maybe there's another problem. good luck, Jenni On Fri, 28 Oct 2005, Martin Ystad wrote: Hi, I'm currently running the latest developmental release for RedHat 9, but I seem to have some problems concerning the intesity-normalization. Unfortunately, my images have some intensity inhomogenities, but the first normalization-step seems to take care of most of these. However, some regions inside the white matter has a somewhat high value, for example 130 -140. These regions are not included in the white matter when I run the recon-all -segmentation program. Adding control-points to these areas does not work, obviously, since only values lower than 110 are considered when the normalization-step is re-runned. (right?) Is there any way to include these areas in my final white matter whithout manually drawing them onto the wm.mgz -volume? For instance raising the threshold for wm-segmentation? If so, would this produce any kind new artifacts which I need to be aware of? Thanks, Martin Ystad Medical Student University of Bergen Institute of Biomedicine Jonas Lies vei 91, 5009 Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer Hi, thanks for your help. I tried what you suggested and it turned out somewhat better. Still, there are large areas that are not included in the wm.mgz. The problem with some of the data sets is that the intensity inhomogenities in the center of the brain also reaches some very low values. I'm worried that if I include these with control-points, some of the gray matter will be misinterpreted as white matter. Is this a legitimate concern, or does the program work in a different way? If so, how low can the control point value be to avoid that this happens? Ps. If I add a control point to one area with a specific high or low value, does this take care of all the other areas with similar values across the brain, or do I have to add a control point to all areas that have discrepancies in intensity? Thanks, Martin ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Normalization problem
I'm working on the latest developmental release for Rh. 9. and I'm having problems with the normalization procedures on my datasets. My datasets have large intensity inhomogeneities due to their acquisition with a surface coil (8-ch. GE). To fix this problem I have resorted to the use of control points. This works if I add enough points across the brain ( over 200), and I get a better normalization result after running "recon-all -normalization -usecontrolpoints" and another skull-stripping on the new T1-volume. To test whether or not this produces a good segmentaition of white matter, I run mri_segment on my brain.mgz -volume. If I'm satisfied with the segmentation result, I move on to the surface processing stage, and run the -autorecon2 -script. The problem is that after running the -autorecon2, the normalization is completely wrong again, looking more like the first normalization done without control points, and the wm.mgz -volume is also bad. So are the surfaces. How do I make freesurfer produce the same good results as I got during the first normalization? Do I need to specify the use of control points again, even though the brain.mgz volume looks ok? Thanks, Martin Ystad Medical Student University of Bergen Institute of Biomedicine Jonas Lies vei 91, 5009 Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Normalization problem
I'm working on the latest developmental release for Rh. 9. and I'm having problems with the normalization procedures on my datasets. My datasets have large intensity inhomogeneities due to their acquisition with a surface coil (8-ch. GE). To fix this problem I have resorted to the use of control points. This works if I add enough points across the brain ( over 200), and I get a better normalization result after running "recon-all -normalization -usecontrolpoints" and another skull-stripping on the new T1-volume. To test whether or not this produces a good segmentaition of white matter, I run mri_segment on my brain.mgz -volume. If I'm satisfied with the segmentation result, I move on to the surface processing stage, and run the -autorecon2 -script. The problem is that after running the -autorecon2, the normalization is completely wrong again, looking more like the first normalization done without control points, and the wm.mgz -volume is also bad. So are the surfaces. How do I make freesurfer produce the same good results as I got during the first normalization? Do I need to specify the use of control points again, even though the brain.mgz volume looks ok? Thanks, Martin Ystad Medical Student University of Bergen Institute of Biomedicine Jonas Lies vei 91, 5009 Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Control-points
Hi, I'm using a lot of control-points (typically 200-500) to compensate for intensity inhomogeneities in my images. This tends to create a rather uniformly looking white matter without any clearly discernible anatomical landmarks, such as the basal ganglia, etc. Does this mean that the subcortical segmentation won't work properly? Will this munificent use of control-points constitute a source of error for the cortical segmentation, or are the protocols relatively robust? Thanks, Martin Ystad Medical Student University of Bergen Institute of Biomedicine Jonas Lies vei 91, 5009 Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] dev-version/new version question.
Hi, I'm currently using the freesurfer-Linux-centos4-dev20060210-full release, and I'm thinking of upgrading to the new stable release. However, I've done a lot of processing with the 20060210-release, and I'm not very keen on starting all over again with the new release. Do you recommend that I reprocess my subjects on the new stable release before doing a group study, or are they similar enough in terms of results, so that I don't need to? In case I do need to run everything over, which steps do I need to rerun (certainly not all?), and can I keep my manual edits to the brainmask- or wm-volume? Thanks, Martin Ystad University of Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Autorecon1
Hi, I have processed some of my subjects on a previously released dev-version (20060210). I have changed to the new stable release. Do I need to run the -autorecon1 step again, or will -autorecon2 and -3 suffice to bring the previously run subjects up to date? (the brainmask-volumes have been edited, so I'd like to keep these changes) Another question; if I run the -autorecon2-wm flag, will I save a lot of computing time if I manually remove the Optic Nerve before doing so? Thanks, Martin Ystad University of Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
[Freesurfer] Talairach question
Hi, I'm doing a group study and I have a question regarding the Talairach transform. I need to make a new transform for some of my subjects using the brainmask.mgz-volume. For others, the original transform is okay (visually) . Is it a problem to have some transforms based on the brainmask volume, and others made from the nu volume? Should I make new transforms for all subjects, even though some of the original ones are fine? Thanks, Martin Ystad University of Bergen, Norway. ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
